OBJECTIVES

: Cancer diagnostic and treatment advances, alongside a growing emphasis on personalised medicine, have contributed to increases in patient and treatment outcome heterogeneity, inherently impacting patient resource use and health utility. The ability to model these relationships is important for accurate economic assessment. This study aims to review considerations of heterogeneity in resource use and health utility quantification in economic analyses of colorectal, lung and ovarian cancer.

METHODS

: A systematic search of published NICE technology appraisals up until April 2020 was conducted, and economic evaluations reviewed. Resource use components and utility estimates were summarised, and the extent to which economic evaluations capture their variation critically reviewed; reported variation between cancers was also described.

RESULTS

: The search identified 49 (colorectal: 9; lung: 31; ovarian: 9) economic analyses. Outcome heterogeneity was captured in utility estimates through stratification by disease progression status (36/49 analyses) or use of time-varying utility functions (12/49 analyses), and in resource use estimates through progression or treatment stratification (43/49 analyses). Mean progression-free utility estimates for colorectal, lung and ovarian cancer were 0.78 (range: 0.73-0.87), 0.70 (0.52-0.82) and 0.76 (0.71-0.83), respectively, and 0.67 (0.60-0.69), 0.53 (0.25-0.78) and 0.70 (0.65-0.76), respectively, for progressed disease. On average, analyses included 6.7 (5-9), 12.2 (1-24) and 2.5 (2-3) resource use components, respectively, with differences across stratifications realised through inclusion of different components or increased component usage. Across all analyses, 29, 65 and 5 unique resource use components were described, respectively. No studies explicitly incorporated patient-level heterogeneity in resource use or health utility estimation.

CONCLUSIONS

: In clinical practice, resource use and health utility are strongly influenced by patient and treatment outcome heterogeneity. However, this study highlights that resource use and health utility are typically modelled based on progression status only, underscoring a need to further explore more granular methods for accurately capturing resource use and health utility in oncology.